Tag
Prism is a semantic document DSL for AI agents that enables structured writing with hierarchy and evidence, rendering as self-contained HTML. It provides 45 production-ready components to replace messy markdown, helping agents produce cleaner, organized outputs.
A detailed blog post dissecting ThunderKittens, a compact DSL for high-performance AI kernels, including a bottom-up analysis of its abstractions and a benchmark implementing a non-causal attention prefill kernel that outperforms FlashAttention-2 by ~1.55x and matches FlashAttention-3.
CPPL is a compiler-mediated framework that bridges LLMs and hardware design by using a Python DSL and JSON-based intermediate representation to enable statically checkable, optimizable RTL generation.
This opinion piece argues that AI kernel portability across different hardware (TPU, GPU, etc.) is structurally impossible due to fundamental hardware differences, and that the best AI stacks will always require hardware-specific DSLs for optimal performance, despite the industry's desire for portability.
Nanopass Framework is a domain-specific language embedded in Scheme for creating compilers through small passes and intermediate representations, reducing boilerplate and improving maintainability.
AgentSPEX introduces a domain-specific language for specifying modular, interpretable LLM-agent workflows with explicit control flow, state management, and a visual editor, outperforming existing Python-coupled frameworks.
TileLang is an open-source domain-specific language and compiler built on TVM for writing high-performance GPU/CPU/NPU kernels such as GEMM and FlashAttention, with recent additions including Huawei Ascend 950 support, an LSP, and multi-backend dialects for CUDA, ROCm, and Metal.